Impact statement
Biocrusts are communities of microorganisms, algae, lichen and mosses that develop in the top few millimeters of the soil. Their poikilohydric condition allows them to become inactive during dry periods and reactivate when water becomes available again, which is crucial in drylands, where water is a limiting factor for vegetation growth. Biocrusts are common in drylands around the world and are important because of the multiple ecological functions they perform. This high tolerance to desiccation suggests that biocrusts can survive long periods of drought. However, some observations seem to indicate changes in biocrust cover during drought periods. In the current context of climate change, knowing biocrust’s limits is essential for the conservation of these areas. This work provides evidence of the impact of increasing drought duration on these communities, analyzing the effects not only on their cover but also on their metabolism (net photosynthesis and dark respiration). On the other hand, analyzing the effects of an increase in precipitation improves our understanding of the limits of biocrust growth. In addition, analyzing various types of biocrusts allows us to better understand the dynamics of the Tabernas Desert’s biocrust communities and provides information for the succession hypothesis.
Introduction
Biocrusts are communities mainly composed of poikilohydric organisms that are capable of surviving in areas where water is a limiting factor. Because their water content tends to equilibrium with that of the environment, they become inactive during dry periods and reactivate when water is available again (Kappen and Valladares, Reference Kappen, Valladares, Pugnaire and Valladares2007; Green et al., Reference Green, Sancho, Pintado, Lüttge, Beck and Bartels2011). This condition makes them particularly important in arid and semi-arid areas, where they protect soil against erosion (Chamizo et al., Reference Chamizo, Cantón, Miralles and Domingo2012; Rodriguez-Caballero et al., Reference Rodríguez-Caballero, Cantón, Chamizo, Lázaro and Escudero2013; Chamizo et al., Reference Chamizo, Rodríguez-Caballero, Román and Cantón2017; Lázaro et al., Reference Lázaro, Gascón and Rubio2023) and can act as main primary producers (Maestre et al., Reference Maestre, Eldridge, Soliveres, Kéfi, Delgado-Baquerizo, Bowker, García-Palacios, Gaitán, Gallardo, Lázaro and Berdugo2016).
Desiccation tolerance mechanisms in these organisms are essential for their survival in these areas because they allow them to maintain their structural and metabolic integrity during dry periods. Some reviews have highlighted the following mechanisms (Kranner et al., Reference Kranner, Beckett, Hochman and Nash2008; Green et al., Reference Green, Sancho, Pintado, Lüttge, Beck and Bartels2011; Heber and Lüttge, Reference Heber, Lüttge, Lüttge, Beck and Bartels2011): (a) compatible solutes that protect membranes and proteins by replacing water molecules during desiccation (Farrar, Reference Farrar1976; Aubert et al., Reference Aubert, Juge, Boisson, Gout and Bligny2007; Oliver, Reference Oliver, Goffinet and Shaw2008; Hoekstra et al., Reference Hoekstra, Golovina and Buitink2001); (b) late embryogenesis abundant proteins (LEAs) and heat shock proteins (HSPs) that protect proteins from denaturation during desiccation (Hoekstra et al., Reference Hoekstra, Golovina and Buitink2001); (c) thermal energy dissipation that protects against photooxidative damage, preventing the formation of reactive oxygen species (ROS) (Heber and Lüttge, Reference Heber, Lüttge, Lüttge, Beck and Bartels2011; Kranner et al., Reference Kranner, Beckett, Hochman and Nash2008); (d) antioxidants that act as scavengers of ROS during desiccation (Kranner et al., Reference Kranner2002; Kranner et al., Reference Kranner, Beckett, Hochman and Nash2008); and (e) amphiphilic metabolites that, although they cause disturbances in the membrane, promote the insertion of antioxidants (Hoekstra and Golovina, Reference Hoekstra and Golovina2002). These mechanisms seem to give biocrusts a certain advantage in dealing with climate change, and they strengthen the belief that biocrusts can survive long periods of drought.
Although several studies have been conducted on climate change in biocrust communities, studies on the effects of changes in precipitation are scarce. Simulations of climate change in southeast Spain found that an increase in soil temperature of 2–3 °C led to losses in biocrust cover because of increased respiration; however, no cover losses were observed with a 30% reduction in precipitation (Maestre et al., Reference Maestre, Escolar, De Guevara, Quero, Lázaro, Delgado‐Baquerizo, Ochoa, Berdugo, Gozalo and Gallardo2013; Ladrón de Guevara et al., Reference Ladrón de Guevara, Lázaro, Quero, Ochoa, Gozalo, Berdugo, Uclés, Escolar and Maestre2014). In contrast, in a similar experiment, 5%–8% losses in moss cover (but not in lichens) were observed because of both increased temperature and reduced precipitation (Li et al., Reference Li, Hui, Zhang and Song2021). On the other hand, research on the effect of altered precipitation patterns by increasing the frequency of small precipitations also revealed losses of biocrust cover related to negative carbon balances (Belnap et al., Reference Belnap, Phillips and Miller2004; Reed et al., Reference Reed, Coe, Sparks, Housman, Zelikova and Belnap2012; Zelikova et al., Reference Zelikova, Housman, Grote, Neher and Belnap2012; Johnson et al., Reference Johnson, Kuske, Carney, Housman, Gallegos‐Graves and Belnap2012). Water availability was shown to increase crust biomass and even to change the crust type (Kidron et al., Reference Kidron, Vonshak, Dor, Barinova and Abeliovich2010). Nevertheless, the effects of prolonged droughts on biocrusts are not very well known, although, droughts drastically affected the crust structure and stability in the Negev (Kidron et al., Reference Kidron, Ying, Starinsky and Herzberg2017). Some research has observed that the duration of droughts can hinder the reactivation of the metabolic activity in lichens and mosses (Munzi et al., Reference Munzi, Varela and Paoli2019; Kranner et al., Reference Kranner, Zorn, Turk, Wornik, Beckett and Batič2003; Harel et al., Reference Harel, Ohad and Kaplan2004; Proctor et al., Reference Proctor, Oliver, Wood, Alpert, Stark, Cleavitt and Mishler2007), while field observations seem to reveal a relationship between periods of drought and cover loss (Belnap et al., Reference Belnap, Phillips and Troxler2006). On the other hand, the effects of increased rainfall on the cover and gas exchange of biocrusts are poorly understood.
The objectives of this work were to analyze the effects of prolonged droughts and increased precipitation on biocrust cover and gas exchange. The state of the photosynthetic systems of the main biocrust components after 3 years of continuous drought was also recorded. We studied five crust types hypothetically representative of successional stages (according to Lázaro et al., Reference Lázaro, Cantón, Solé-Benet, Bevan, Alexander, Sancho and Puigdefábregas2008 and Rubio and Lázaro, Reference Rubio and Lázaro2023, among others). We hypothesized that (a) episodes of hydration insufficient to produce positive net photosynthesis would lead to a decrease in biocrust biomass due to respiration and biocrust cover would be visibly reduced after a few years of drought; and (b) an increase in precipitation would lead to an increase in net photosynthesis rates promoting the growth of biocrust, which could visibly increase its cover.
Material and methods
Study area
This study was performed at the El Cautivo field site in the Tabernas Desert (Almería, Spain). This widely studied area (Alexander et al., Reference Alexander, Harvey, Calvo-Cases, James, Cerdá, Millington and Pye1994; Raggio et al., Reference Raggio, Pintado, Vivas, Sancho, Büdel, Colesie, Weber, Schroeter, Lázaro and Green2014; Miralles et al., Reference Miralles, Lázaro, Sánchez-Marañón, Soriano and Ortega2020, among others) has a mean annual rainfall of about 230 mm, distributed mainly during autumn and winter, and a mean annual temperature of 18 °C, which can reach 45 °C in summer and −5.5 °C in winter (Lázaro et al., Reference Lázaro, Rodrigo, Gutiérrez, Domingo and Puigdefábregas2001, Reference Lázaro and Pandalai2004). The weathering of marine marls from the Upper Miocene has generated a landscape of badlands, heterogeneously colonized by vascular vegetation and biocrusts (Lázaro et al., Reference Lázaro, Cantón, Solé-Benet, Bevan, Alexander, Sancho and Puigdefábregas2008). The vegetation is patchy, concentrated in certain landforms and is mainly made up of three biotypes: tussock grasses, dwarf shrubs and annual herbs. Vegetation covers approximately one-third of the territory, while another third features eroded regolith with hardly any vegetation and the final third is covered with biocrusts, which are also in the plant interspaces.
Experimental design
Five crust types characteristic of the area were analyzed, which could represent stages of a hypothetical succession (Lázaro et al., Reference Lázaro, Cantón, Solé-Benet, Bevan, Alexander, Sancho and Puigdefábregas2008; Rubio and Lázaro, Reference Rubio and Lázaro2023); ordered from earliest to latest, they were as follows:
-
- Physical crust (P): Bare soil with a low amount of microorganisms not visible to the naked eye in the field. Smooth surface and beige, pale grey or whitish color.
-
- Incipient cyanobacterial biocrust (I): Located in flat and sun-exposed areas undergoing relatively frequent trampling. It has a slight bacterial colonization. The five main phyla are: Proteobacteria (14.8%), Bacteriodetes (14.6%), Actinobacteria, (14.4%), Cyanobacteria (12.4%) and Chloroflexi (11.3%) (Miralles et al., Reference Miralles, Lázaro, Sánchez-Marañón, Soriano and Ortega2020). Smooth compact surface and pale brown or yellowish color.
-
- Mature cyanobacterial biocrust (C): A widespread biocrust becoming dominant on sun-exposed areas, with a higher microbial concentration than I (Miralles et al., Reference Miralles, Lázaro, Sánchez-Marañón, Soriano and Ortega2020). The five main phyla are: Cyanobacteria (21.9%), Bacteriodetes (14.3%), Proteobacteria (13.2%), Actinobacteria (9.8%) and Chloroflexi (9.7%) (Miralles et al., Reference Miralles, Lázaro, Sánchez-Marañón, Soriano and Ortega2020). Büdel et al. (Reference Büdel, Colesie, Green, Grube, Suau, Loewen‐Schneider, Maier, Peer, Pintado, Raggio, Ruprecht, Sancho, Schroeter, Türk, Weber, Wedin, Westberg, Williams and Zheng2014) found 14 cyanobacterial genera, highlighting Nostoc, Leptolyngbya, Scytonema, and Phormidium. Some filamentous cyanobacteria have been identified to the species level, such as the heterocystous Tolypothrix distorta and Scytonema hyalinum and the non‐heterocystous Leptolyngbya frigida, Microcoleus steenstrupii and Trichocoleus desertorum (Roncero-Ramos et al., Reference Roncero‐Ramos, Muñoz‐Martín, Chamizo, Fernández‐Valvuena, Mendoza, Perona, Cantón and Mateo2019). Rough surface and brown color. Some small pioneer lichens such as Fulgensia desertorum Poelt, Fulgensia poeltii Llimona and Endocarpon pussillum Hedw are often present.
-
- Lichen biocrust dominated by Squamarina lentigera Poelt and/or Diploschistes diacapsis Lumbsch (S): It is the most widespread biocrust type at the field site, occupying mainly north and east-oriented hillslopes. Rough surface and whitish color. It includes a diversity of lichens, such as Buellia zohary Galun, Diploschistes ocellatus Llimona, and Psora decipiens Hoff.
-
- Lichen biocrust characterized by Lepraria isidiata Llimona & Crespo (L). It develops on the shadiest north-facing hillslopes, often in spaces among plants (which show 20%–40% cover). Others lichens such as Squamarina cartilaginea P. James, Xanthoparmelia pokornyi Blanco, Crespo, Elix, Hawksw. & Lumbsch, and Teloschistes lacunosus Savicz, as well as mosses such as Grimmia pulvinata Sm, are also characteristic. Rough surface and a mosaic of whitish (often dominant), green and dark colors.
We selected two representative areas per crust type, and nine representative plots were established in each area, delimited by 10-cm-diameter transparent methacrylate rings, distributed in three sets of three samples. We used six replicates per treatment and crust type; every set of plots included the following three treatments:
-
- Control (C): samples exposed to the natural rainfall regime.
-
- Watering (W): samples subjected to irrigation, doubling the natural rainfall. To do so, we measured precipitation and, the day or days following each precipitation event, the samples were carefully watered with an amount of demineralized water equal to that received by precipitation. Watering turned out to be a slow process, lasting several days in large events and when gas exchange measurements were imminent (see below), in which case the order of the irrigations was used to order the measurements so that they were done under similar moisture conditions.
-
- Rain exclusion (RE): samples permanently covered with a square, 20-cm-side, transparent methacrylate roof at a height of approximately 20 cm from the ground, preventing the entry of rainwater but allowing light to enter. Although these rainout shelters often condense water, they probably barely decrease the high relative humidity associated with rain, so they do not exactly replicate the conditions of natural drought.
CO2 fluxes measurement
The net photosynthesis and dark respiration were measured periodically using an open-circuit infrared gas analyzer LI-6400 (Lincoln, USA) connected to a transparent chamber of 668 cm3 designed and calibrated by Ladrón de Guevara et al. (Reference Ladrón de Guevara, Lázaro, Quero, Chamizo and Domingo2015). To measure respiration, the chamber was covered with an opaque cloth and a new record was taken immediately after each light measurement. For photosynthesis and respiration and for each plot in each campaign, one record consisted of the average of five consecutive measurements. Ten campaigns were carried out over 3 years of experimentation: two annual (winter and summer) and some additional ones after heavy rainfall events. Because biocrust activity peaks during the first light hours of the morning (Raggio et al., Reference Raggio, Pintado, Vivas, Sancho, Büdel, Colesie, Weber, Schroeter, Lázaro and Green2014; Ladrón de Guevara et al., Reference Ladrón de Guevara, Lázaro, Quero, Ochoa, Gozalo, Berdugo, Uclés, Escolar and Maestre2014), only one crust type could be measured per day, so every campaign required five consecutive sunny days. We staggered the irrigations so that each day we measured the crust type watered the day before. We considered positive values to be CO2 consumption by the biocrust and negative values to be CO2 loss by the biocrust.
Cover estimation
Each plot was photographed twice per year (winter and summer) to obtain the cover of the main components of the biocrust (bare soil, cyanobacteria, lichens and mosses). Covers were approximated from the frequencies, using the program GIMP 2.10.34 (GIMP Development Team, 2023) to draw a regular 196-cell grid and to overlap it over each plot photograph. Because 59 grid cells fell outside the plot ring, the frequencies were counted on 137 cells. Frequency counts have been used as a cover subrogate (Maestre et al., Reference Maestre, Escolar, De Guevara, Quero, Lázaro, Delgado‐Baquerizo, Ochoa, Berdugo, Gozalo and Gallardo2013) because it is an objective and repeatable method. This method tends to overestimate the cover of the small-thalli species; however, distinguishing only a few cover categories, the error can be assumed to be acceptable. We could not monitor the chlorophyll to avoid disturbing the small plots through successive sampling, but the small plot size and the grid allowed us to visually check the quality of the cover estimation.
Measurements of chlorophyll a fluorescence
To analyze the effect of treatments on photosystem II (PSII) efficiency and to determine whether each measured organism was alive, we measured the chlorophyll a fluorescence of each sample using a MINI-PAM Photosynthesis Yield Analyzer (Heinz Walz GmbH, Germany). Measurements were performed at night and, 30 min before measuring, the samples were sprayed with demineralized water until saturation was reached in the surface horizon (about 4 mm). We distinguished physical crust, incipient and mature cyanobacteria, mosses and the main lichen species. A hard grid of 11 × 12 cm with 99 cells was superimposed on each ring, always in the same positions to ensure measurements were taken at the same points or on the same thalli across the dates. A variable number of measurement points were selected to represent each sample, depending on the diversity and abundance of organisms present. We measured species and crust types that appeared in at least three cells to obtain three replicates per surface category and sample. Therefore, some plots only had three measurements, for example, those in which there was nothing more than physical crust or incipient cyanobacterial biocrust, whereas other samples were the subjects of many measurements. All measurements obtained in every plot were used in the analyses; for graphical representation, we used a single (average) fluorescence value per plot. We carried out three measurement campaigns: March 2020, October 2020 and January 2021.
Climate data
Data on precipitation were obtained from five climate stations installed in one of the two zones for each crust type, measured by Rain-O-Matic-Pro tipping-bucket rain gauges of 0.25-mm resolution (Pronamic, Denmark). The missing data were filled in using the data from the nearest rain gauge based on the regression between the two rainfall data series. For every month, we calculated the amount of precipitation and the number of rainfall events defined by a minimum inter-event time of 6 h.
Data analysis
To test for differences, for each plot, the covers of bare soil, cyanobacteria, lichens and mosses, as well as the net photosynthesis and dark respiration, were analyzed as dependent variables using generalized mixed models (GLMM), assuming that data follows a gamma log link distribution. Years (in the case of cover) or times of measurement (in the cases of net photosynthesis and dark respiration) were considered within-subject factors and crust type and treatment were between-subject factors in the three cases, and their interactions were analyzed.
On the other hand, chlorophyll a fluorescence, including all the measurements at the points selected in each sample, was analyzed using generalized models (GLMs), assuming that data follows a gamma log link distribution. Crust type and treatment were considered factors, and their interaction was also analyzed. For all interactions, multiple comparisons were analyzed using the Bonferroni test.
All the analyses were made using SPSS 28.0 (IBM Corporation, USA). Differences were assumed significant at p < 0.05.
Results
Precipitation
The averages of total annual precipitation for the years 2018–2021 were 217.80, 212.44, 151.14 and 312.78 mm, respectively.
Changes in cover
After the 3-year experiment, treatment and crust type significantly affected the cover of all the components, except for moss (Table 1). Time itself affected the cover of cyanobacteria, mosses and (indirectly) the bare soil. However, treatments interacted with time (except in moss cover) and with the type of crust in lichen cover, where the three factors interacted significantly, affecting the bare soil cover.
Table 1. F-values in the results of GLMM analyses for cover of biocrust principal components (bare soil, cyanobacteria, lichens and mosses) and metabolism rates (net photosynthesis and dark respiration), and results of GLM analyses for chlorophyll a fluorescence
Note: * indicates the significant effects (p-value < .05).
Bare soil increased significantly under the rain exclusion treatment in all biocrusts (Figure 1), increasing by 23% in Incipient, 52% in Cyanobacteria, 31% in Squamarina and 21% in Lepraria crusts. Nevertheless, watering only significantly affected Cyanobacteria and Squamarina, reducing bare soil cover by 9% and 4%, respectively.
Figure 1. Cover of bare soil, cyanobacteria, lichens and mosses of each treatment, at the beginning (2018) and at the end (2021) of the experiment in each of the crust types (Physical, Incipient, Cyanobacteria, Squamarina and Lepraria). The bars represent the averages and error bars represent the average ± 95% confidence level. *indicates significant differences (p < .05) based on the Bonferroni test.
Cyanobacterial cover only changed significantly under the rain exclusion condition in Physical and Cyanobacteria crusts, where it decreased by 23% and 48%, respectively (Figure 1). The increase in cyanobacterial cover promoted by watering was not significant. Cyanobacterial cover was not significantly affected by treatments or time in the lichen-dominated crusts (Figure 1).
Lichen cover decreased significantly with rain exclusion in the Incipient and Cyanobacteria crusts (virtually disappearing in both) and in the Squamarina crust (where it decreased by 29%; however, its decline in the Lepraria crust was not significant (Figure 1). Note that the reduction of lichen cover in control plots of Lepraria was almost 15%. Watering only significantly affected lichen cover in the Incipient crust, where it increased by 5%.
Finally, moss cover (identified only in Squamarina and Lepraria) decreased significantly over time, irrespective of the treatment and crust types (Table 1 and Figure 1).
Changes in CO2 fluxes
The three factors, treatment, crust type and time, significantly affected net photosynthesis and dark respiration. The interactions of treatments with crust type and time, as well as the triple interaction of the three factors, were significant for both dependent variables (Table 1). Overall, net photosynthesis was mostly negative and was positive only in September 2019, December 2019 and January 2020 under the control and watering treatments (Figure 2).
Figure 2. Evolution of net photosynthesis (with regression lines) per treatment and crust type in relation to the rainfall from the climate station representative of each crust type. Symbols represent the averages, and error bars represent the average ± 95% confidence level. In the legend of each graph at the upper left corner, C means control, W means watering treatment and RE means rain exclusion. The superscript letters in these treatment symbols indicate whether the differences between treatments are significant (two treatments are different if they do not share any letters).
In the Physical crust, the treatments did not significantly change photosynthesis or respiration after 3 years (Figures 2 and 3).
Figure 3. Evolution of dark respiration (with regression lines) per treatment and crust type in relation to the rainfall from the climate station representative of each crust type. Symbols represent the averages, and error bars represent the average ± 95% confidence level. In the legend of each graph at the upper left corner, C means control, W means watering treatment and RE means rain exclusion. The superscript letters in these treatment symbols indicate whether the differences between treatments are significant (two treatments are different if they do not share any letters).
In the Incipient crust, net photosynthesis increased significantly under the watering treatment (Figure 2) from −0.19 μmol/m2s in March 2018 to −0.08 μmol/m2s in January 2021. Punctual increases observed in September 2018 (0.21 μmol/m2s), December 2019 (0.24 μmol/m2s) and January 2020 (0.39 μmol/m2s) coincided with rain events (Figure 2). Dark respiration was significantly higher in the watering treatment (Figure 3).
In the Cyanobacteria crust, net photosynthesis was not significantly different among the treatments (Figure 2) although it increased over the 3 years by 0.5 μmol/m2s under the watering treatment. As in the case of the Incipient crust, the positive rates of net photosynthesis reached in December 2019, January 2020 and January 2021 coincided with rain events. Dark respiration was significantly higher under the watering treatment and significantly lower under rain exclusion (Figure 3).
In the Squamarina crust, net photosynthesis was significantly higher in watering conditions than in rain exclusion, reaching 0.23 μmol/m2s in January 2021 (Figure 2). Increased photosynthesis was observed in the control and watering samples in January 2020 and January 2021, coinciding with rain events. Dark respiration was significantly higher under the watering treatment, increasing by 0.37 μmol/m2s and significantly lower under rain exclusion (Figure 3). In September 2020, dark respiration of watering treatment increased to −0.98 μmol/m2s with the first rainfall after summer drought.
In the Lepraria crust, net photosynthesis under watering was not significantly different from that of the control but did differ from that under rain exclusion (Figure 2). Conversely, dark respiration was significantly higher with watering (thus making it difficult the increase of net photosynthesis) but it was not significantly lower under exclusion, and it peaked at −1.31 μmol/m2s by September 2020 with the first rainfall after summer (Figure 3).
Chlorophyll a fluorescence
Chlorophyll a fluorescence was significantly different among crusts and treatments and the effect of the treatments depended on the crust type (Table 1). Fluorescence was significantly lower under rain exclusion than in control or watering treatment in Incipient, Squamarina and Lepraria crusts, while watering did not produce a difference with respect to the control (Figure 4). Fluorescence in lichenic crusts was significantly higher than that of cyanobacterial crusts in both control and watering. Rain exclusion affected fluorescence more in Lepraria than in Squamarina.
Figure 4. Comparison of chlorophyll a fluorescence of each treatment in each crust type. Symbols represent the averages and error bars represent the average ± 95% confidence level. P, Physical crust; I, Incipient crust; C, Cyanobacteria crust; S, Squamarina crust; L, Lepraria crust. In the legend at the upper right corner, C means control, W means watering treatment and RE means rain exclusion.
Discussion
Our treatments significantly affected the cover and metabolism of biocrusts, with different effects depending on the biocrust types assumed to be successional stages.
We achieved the maximum possible replication of CO2 measurements considering that the daily metabolic cycle (Ladrón de Guevara et al., Reference Ladrón de Guevara, Lázaro, Quero, Ochoa, Gozalo, Berdugo, Uclés, Escolar and Maestre2014) affects the fluxes if each round of measurements lasts too long, depending on the number of plots. However, the punctual nature of the gas exchange measurements, the seasonality, the erratic nature of rainfall, the inevitable increase in the number of events by watering and the necessity to water on two or more successive days after the major rains or for the gas exchange campaigns surely added noise to the data on net photosynthesis and respiration, blurring their relationships with the factors and the cover variations.
Effect of prolonged droughts
Rain exclusion negatively and differentially affected both the cover and metabolism of all biocrusts. The Cyanobacteria crust lost more than 40% of its cyanobacterial cover, whereas the Incipient lost 20% (Figure 1). Net photosynthesis did not decrease significantly in either (Figure 2); however, dark respiration was significantly lower in the Cyanobacterial crust (Figure 3). Lichen-dominated crusts lost 30% and almost 20% of lichen cover in Squamarina and Lepraria crusts, respectively (Figure 1). However, unlike in the Cyanobacteria crust, they did not lose the cyanobacterial cover. This suggests that prolonged droughts (and possibly other disturbances), can reverse the direction of succession. Although lichens displace cyanobacteria under favorable conditions (Lázaro et al., Reference Lázaro, Cantón, Solé-Benet, Bevan, Alexander, Sancho and Puigdefábregas2008), when lichens recede, cyanobacteria occupy their space. This is the only explanation for the fact that cyanobacterial cover decreased where Cyanobacteria dominate but not where lichens dominate. The space left by the lichen retreat would have particularly suitable conditions for cyanobacteria despite the drought due to the physical and chemical changes lichens produce in soil. These changes include improvement of soil structure, porosity, stability, water retention and accumulation of fine-grained material (Miralles et al., Reference Miralles, Cantón and Solé-Benet2011, Chamizo et al., Reference Chamizo, Cantón, Miralles and Domingo2012, Reference Chamizo, Belnap, Eldridge, Canton, Malam Isa and Weber2016); along with increased organic carbon, nitrogen and nutrients: Belnap and Eldridge (Reference Belnap, Eldridge, Belnap and Lange2003) showed that Carbon and Nitrogen are fertilizers that increase the amount of chlorophyll a in cyanobacteria. This is consistent with the results of Zelikova et al. (Reference Zelikova, Housman, Grote, Neher and Belnap2012), who observed an increase in cyanobacterial cover associated with a decrease in moss cover.
Maestre et al. (Reference Maestre, Escolar, De Guevara, Quero, Lázaro, Delgado‐Baquerizo, Ochoa, Berdugo, Gozalo and Gallardo2013) found that biocrust cover and metabolism were significantly affected by a 2–3 °C temperature increase, but not by a 30% reduction in precipitation although, as Ladron de Guevara et al. (Reference Ladrón de Guevara, Lázaro, Quero, Ochoa, Gozalo, Berdugo, Uclés, Escolar and Maestre2014) observed, the open-top-chambers used to increase temperature probably caused a decrease in dew, fog and rain. Non-rainfall water inputs (NRWI) can be relevant for biocrust activity in drylands (del Prado and Sancho, Reference Del Prado and Sancho2007; McHugh et al., Reference McHugh, Morrissey, Reed, Hungate and Schwartz2015). However, to date, there is not enough reliable NRWI data from our study area. Our highest net photosynthesis rates tended to coincide with rainfall periods (Figure 2). Therefore, a 30% reduction in precipitation may not be sufficient to observe short-term decreases in cover. This has not been widely studied in biocrust, but Miranda et al. (Reference Miranda, Padilla, Lázaro and Pugnaire2009) found that a 25% reduction in precipitation did not significantly affect plant cover in the short term in the same area; however, a 50% reduction did. Desiccation tolerance is species-specific (Green et al., Reference Green, Sancho, Pintado, Lüttge, Beck and Bartels2011). We have not found data on the desiccation tolerance of our main species, but many organisms are able to survive in a latent state for drought periods longer than 3 years (Alpert, Reference Alpert2000). These organisms would not necessarily die or lose cover visibly under our experimental drought.
Effect of increased precipitation
The watering did not significantly increase biocrust cover over the short term (Figure 1). However, it increased net photosynthesis in all crusts except in Lepraria (although that increment was not significant in Cyanobacteria). Additionally, watering increased dark respiration in all, but particularly in lichen-dominated crusts (Figures 2 and 3). This is consistent with Lange (Reference Lange1980). The short periods of positive net photosynthesis in Cyanobacteria agree with Büdel et al. (Reference Büdel, Williams and Reichenberger2018), who explained that the metabolic active period commences with up to 3 months of carbon loss, likely due to the reestablishment of the structures of the organisms, prior to about a 4-month period of net carbon gain. In the Tabernas Desert, the period of net carbon gain seems to be still shorter than in the Australian Gulf Savannah.
The low increase in cover despite the increase in metabolism could be influenced by the small plot size (Rubio and Lázaro, Reference Rubio and Lazaro2024). However, the control samples did not show significant changes in cover (Figure 1) and it is unlikely that the ring affected only the watering samples. Therefore, we propose two not-exclusive explanations for the low cover increase with watering. (a) The experiment only lasted 3 years and, although cyanobacteria can grow rapidly, lichens and mosses develop more slowly (Dojani et al., Reference Dojani, Büdel, Deutschewitz and Weber2011; Rubio and Lázaro, Reference Rubio and Lázaro2023), and their growth rate depends on the species and the environment (Belnap and Eldrige, 2001; Weber et al., Reference Weber, Bowker, Zhang, Belnap, Weber, Büdel and Belnap2016); (b) Since we selected the plots based on their representativeness, they had low bare-soil cover. Therefore, an increase in the cover of a surface category had to occur mainly at the expense of the cover of another category and it is unlikely that the competition between lichen and cyanobacterial covers would be resolved in such a short time.
Although we altered the annual timing of rainfall as little as possible by watering after each natural rain event, we could not avoid doubling the number of rainy days in the plots under watering. Regardless of the total amount of precipitation, changes in precipitation patterns can decrease CO2 fluxes and cover in plants (Knapp et al., Reference Knapp, Fay, Blair, Collins, Smith, Carlisle, Harper, Danner, Lett and McCarron2002) and the increased frequency of small rainfall events significantly decreased biocrust cover (Belnap et al., Reference Belnap, Phillips and Miller2004; Reed et al., Reference Reed, Coe, Sparks, Housman, Zelikova and Belnap2012) because with increased frequency each precipitation event is less abundant, facilitating water evaporation (Munzi et al., Reference Munzi, Varela and Paoli2019). This can force biocrusts to remain in a desiccated state (Williams et al., Reference Williams, Büdel, Reichenberger and Rose2014; Kranner et al., Reference Kranner, Zorn, Turk, Wornik, Beckett and Batič2003; Proctor et al., Reference Proctor, Oliver, Wood, Alpert, Stark, Cleavitt and Mishler2007) and occasionally even to die (Reed et al., Reference Reed, Coe, Sparks, Housman, Zelikova and Belnap2012). However, in this case an increase in the number of rain days does not mean a decrease in rainfall volume per day. Moreover, this negative effect of the increased frequency of small events contrasts with the hypothesis of Lázaro et al. (Reference Lázaro, Rodrigo, Gutiérrez, Domingo and Puigdefábregas2001) and Lázaro (Reference Lázaro and Pandalai2004) from our study area, suggesting that the higher frequency of small rainfall events with regard to the surrounding areas would selectively benefit biocrusts over vascular vegetation, explaining the abundance of biocrusts in this area. Nevertheless, both hypotheses are not mutually exclusive. Belnap et al. (Reference Belnap, Phillips and Miller2004) and Reed et al. (Reference Reed, Coe, Sparks, Housman, Zelikova and Belnap2012) investigated in the southwestern USA, where rainfall occurs mainly in summer, when water can evaporate quickly. In the southeast of Spain, rainfall occurs mainly in autumn and winter, when temperatures are lower and the soil remains wet for longer.
The succession and the response to changes in precipitation
The successional hypothesis is widely accepted worldwide (Belnap and Eldridge, Reference Belnap, Eldridge, Belnap and Lange2003; Büdel et al., Reference Büdel, Darienko, Deutschewitz, Dojani, Friedl, Mohr and Weber2009; Zhuang et al., Reference Zhuang, Zhang, Zhao, Wu, Chen and Zhang2009; Drahorad et al., Reference Drahorad, Steckenmesser, Henningsen, Lichner and Rodný2013; Geng et al., Reference Geng, Zhou, Wang, Peng, Li and Li2024), although not unanimously: Kidron (Reference Kidron2019) and Kidron and Xiao (Reference Kidron and Xiao2024) claimed that succession can only be invoked when the successive communities in a recovery space are compared to surrounding ones. The successional hypothesis has been widely assumed in the Tabernas Desert (Lázaro et al., Reference Lázaro, Cantón, Solé-Benet, Bevan, Alexander, Sancho and Puigdefábregas2008; Chamizo et al., Reference Chamizo, Rodríguez-Caballero, Cantón, Asensio and Domingo2015; Miralles et al., Reference Miralles, Lázaro, Sánchez-Marañón, Soriano and Ortega2020; Lopez-Canfin et al., Reference Lopez-Canfin, Lázaro and Sánchez-Cañete2022a, Reference Lopez-Canfin, Lázaro and Sánchez-Cañete2022b; Rubio and Lázaro, Reference Rubio and Lázaro2023; among others). Our results according to the crust type support this hypothesis; the successional order in the Tabernas Desert would be Physical, Incipient, Cyanobacteria, Squamarina and Lepraria (Lázaro et al., Reference Lázaro, Gascón and Rubio2023).
Incipient crust did not show significant cover losses under rain exclusion, maybe because its relatively frequent trampling provides it with greater adaptation to disturbances. Furthermore, Incipient’s net photosynthesis and respiration rates were like those of the Physical crust and lower than those of the Cyanobacterial crust. These differences, along with the lower microbial biomass of Incipient (Miralles et al., Reference Miralles, Lázaro, Sánchez-Marañón, Soriano and Ortega2020), its visibility to the naked eye in the field and its persistence over time in trampled places, foster the consideration of Incipient as a successional stage between the Physical and the Cyanobacterial crusts. Rain exclusion caused a decreasing cover loss through the pothetical succession (50%, 30% and 20% in Cyanobacteria, Squamarina and Lepraria crusts, respectively). The early successional stages might be more sensitive to environmental changes because the later ones involve higher biodiversity and, therefore, higher functional redundancy, thus achieving greater resilience (Biggs et al., Reference Biggs, Yeager, Bolser, Bonsell, Dichiera, Hou, Keyser, Khursigara, Lu, Muth, Negrete and Erisman2020). This growing resistance to cover loss is consistent with the fact that ecosystem services increase accompanying biocrust succession, according to various empirical findings, such as decreasing erodibility along succession (Lázaro et al., Reference Lázaro, Gascón and Rubio2023), increasing water collection and retention (Chamizo et al., Reference Chamizo, Belnap, Eldridge, Canton, Malam Isa and Weber2016) and growing nutrient accumulation (Zhang et al., Reference Zhang, Gao, Yu, Zhang, Yan, Wu, Song and Li2022).
Ongoing climate change could reduce biocrust cover; however, this will not necessarily occur because climate change is slower than simulated in experiments and affects several generations of organisms, giving species the opportunity to acclimatize and even adapt (Pintado et al., Reference Pintado, Sancho, Green, Blanquer and Lázaro2005). The models indicate a progressive concentration and intensification of precipitation, with lengthening droughts (IPCC, Reference Lee and Romero2023). However, protecting at least the lichen-dominated biocrusts -the best at resisting droughts and providing ecosystem services- is crucial because, although biocrusts can resist high rainfall intensity (Lázaro et al., Reference Lázaro, Gascón and Rubio2023), they are at serious risk with the intensification of land use. Moreover, we have verified here that prolonged droughts indeed increase periods of negative carbon balance. On the other hand, a hypothetical substantial increase in natural precipitation would benefit vascular plants rather than biocrusts (Lázaro, Reference Lázaro, Rodriguez-Tamayo, Ordiales, Puigdefábregas, Moya, Cabello, Cerillo and Rodriguez-Tamayo2004). This area is currently below the forest’s lower climatic limit; with double the current rainfall, it would exceed this limit and the vegetation would barely leave room for biocrusts.
Conclusions
Our results confirm our initial hypothesis. Although biocrusts can survive long periods of drought, prolonged droughts reduced biocrust biomass by decreasing opportunities to reach the compensation point and accumulate carbon by net photosynthesis, whereas they increased periods of negative carbon balance, potentially causing a significant cover loss. However, this decline in cover will not necessarily occur in the long term because many of these species acclimatize or adapt, as their geographic distributions show. In the experiments, instant artificial climatic changes are often applied to specific individuals, which, in the case of lichens, are sets of symbionts that have organized and developed slowly in equilibrium with the environmental conditions that prevailed until the time of the experiment; so, the results could overestimate the effects of climate change. Moreover, climate change models do not predict years-long droughts in this region.
Our results show that increased rainfall would not necessarily imply increased biocrust cover in the long term because the current biocrust cover is already high; additionally, an increase in precipitation would favor the development of vascular vegetation, which ultimately would outcompete biocrusts, as can be observed in the regions surrounding the Tabernas Desert. This strongly suggests that the current precipitation conditions in the Tabernas Desert are close to being optimal for biocrusts to reach their maximum possible extension in the area.
Our results also show that biocrust’s response to changes in precipitation depends on the biocrust type. Therefore, the analysis of these various responses is essential to better understand biocrust dynamics and the associated processes, as well as for issues of land management and conservation. Our results support the succession hypothesis. The development of lichens increases community resilience. Changes in biocrust cover as a consequence of rainfall changes appear muffled in the biocrust community hypothetically considered late-successional. Therefore, our results suggest that the dynamic relationships among biocrust types should be considered in future work at other field sites.
Open peer review
For open peer review materials, please visit https://doi.org/10.1017/dry.2025.5
Data availability statement
The data that support the findings of this study are available from the corresponding author, R.L., upon reasonable request.
Acknowledgments
We thank the Viciana brothers, landowners of El Cautivo, who kindly allowed us to conduct this study and Maria Daskalaki for her help in irrigating the samples.
Author contribution
All authors contributed to the conception and design of the study. Roberto Lázaro and Clement López selected the plots and prepared the material, while all authors carried out data collection and analysis. Consuelo Rubio wrote the first draft of the manuscript and Roberto Lázaro and Clement López commented on later versions of the manuscript. All authors read and approved the final manuscript.
Financial support
This work was supported by the Ministry of Science, Innovation and University (R.L., Grant No. CGL2016–78075-P), (J.L.G-R, Grant No. PID2020-117825GB-C21), (D.F, Grant No. PID2020-117825GB-C22) and (C.R., Grant No. FPU18/00035); the Andalusian Plan for Research, Development and Innovation through the project BAGAMET (F.P, Grant No. P20_00016); and European Union’s Horizon (C.L., Marie Skłodowska-Curie grant agreement Nº 101109110).
Competing interests
The authors declare none.
Open access
Comments
No accompanying comment.